Mantis Bugtracker

Viewing Issue Simple Details Jump to Notes ] View Advanced ] Issue History ] Print ]
ID Category Severity Reproducibility Date Submitted Last Update
0000306 [Yate - Yet Another Telephony Engine] module crash always 2012-06-12 01:18 2013-05-21 15:20
Reporter vankooch View Status public  
Assigned To
Priority normal Resolution open  
Status feedback   Product Version 4.x
Summary 0000306: Callfork module does not work correct
Description Hi,

after days of searching I finally found the trouble module. We had strange behaviors of yate:

- engine timer was not executed (register.conf) on very small load 1-2 cps
- rmanager lags. hangs and crashes on status command
- yate crash on loads of 6-10cps

I can reproduce the behavior on ubuntu 10.04 and 11.10 in vm and on bare metal with yate 3.3.3 rev. 4485 and yate 4.1.1
It feels like that yate 3.3.3 reacts a bit better... but it's only a feeling ;)

We use h323 and sip module and routing is made with register with postgresql backend, we also use copyparam in preroute and route. The db returns one row with 10 callto legs and paramters for each:

"callto.1.caller",
"callto.1.callername",
"callto.1.format",
"callto.1.formats",
"callto.1.line",
"callto.1",
"callto.1.maxcall",
"callto.1.osip_P-Asserted-Identity",
"callto.1.osip_Gateway-ID",
"callto.1.osip_Tracking-ID",
"callto.1.rtp_addr",
"callto.1.rtp_forward",
"callto.1.rtp_localip",
"callto.1.rtp_port",
"callto.1.timeout",
"callto.1.scgatewayid",
"callto.1.scgatewayaccountid",
"callto.1.scgatewayipid",
"callto.1.sctechcalled",

Most times only the first 4 callto legs are filled with informations,callto.2 with |next=5000 and callto.4 with |drop=25000, the rest is NULL.

Please let me know if you need more informations or I should check thinks out.

Thank you!
Additional Information
Tags No tags attached.
Attached Files

- Relationships

-  Notes
(0000524)
paulc (administrator)
2013-04-19 13:43

Is it possible to retest with a more recent version of Yate?
(0000532)
vankooch (reporter)
2013-05-21 15:20

Hi,

I've tested with Yate 4.3.1 r5466 and I still have that problem.
Yate has routed 56 call to an h323 gateway, then it hangs / crashed.

From Logs I only get this informations:

20130521114327.376447 <FAIL> Peer ID failed - timeout on call endpoint mutex owned by 'Call Router'!
20130521114331.585317 <ALL> Cleaning up RTP 0x7fa1c0019d80 [0x7fa1c0019b80]
20130521114332.376639 <FAIL> Call disconnect failed - timeout on call endpoint mutex owned by 'Call Router'!
20130521114336.585486 <FAIL> Peer ID failed - timeout on call endpoint mutex owned by 'Call Router'!


So let me know if I need to retest.

- Issue History
Date Modified Username Field Change
2012-06-12 01:18 vankooch New Issue
2013-04-19 13:43 paulc Note Added: 0000524
2013-04-19 13:43 paulc Status new => feedback
2013-05-21 15:20 vankooch Note Added: 0000532


Copyright © 2000 - 2008 Mantis Group
Powered by Mantis Bugtracker